Modeling SuperKEKB Backgrounds with the Belle II Electromagnetic Calorimeter

The largest current obstacle to SuperKEKB’s luminosity goals is currently beam-related backgrounds occurring during accelerator operation. Thus, understanding the level of these backgrounds is of crucial importance for the future of the facility. In this work, we take advantage of the Belle II Electromagnetic Calorimeter’s near-total coverage of the interaction region to create a spatial model of beam-induced backgrounds with the aim of providing fast feedback to improve accelerator conditions.


Introduction
The SuperKEKB accelerator [1] is at the forefront of intensity frontier physics, and aims to increase its luminosity to the order of 6 × 10 35 cm −1 s −1 by increasing the beam currents while simultaneously reducing beam sizes at the interaction point, in what is known as the "nanobeam" scheme.As beam luminosity increases, however, the rate of beam-related backgrounds increase correspondingly, creating an increasingly challenging environment.Beam backgrounds incident on the Belle II detectors cause undesired background signals with the data, interfere with detector operation, and cause accelerated decay of sensitive detector components, and thus the reduction of these events is of critical importance for SuperKEKB and Belle II to achieve their goals.
Previous studies of Belle II backgrounds during beam commissioning [2,3] have relied on dedicated detectors designed to measure specific backgrounds, along with input from selected Belle II physics detectors, deployed with a relatively small coverage.In this study, we consider the Belle II Electromagnetic Calorimeter (ECL) [4] to fit beam-induced backgrounds to a spatial distribution, making use of the ECL's near-total coverage of the interaction region.
Beam-induced backgrounds result from several distinct phenomena, each of which scales differently with beam conditions.In this study, we consider mainly beam-induced backgrounds from two sources: Touschek and beam-gas backgrounds.Other background sources such as synchrotron radiation are better measured with other detectors and are discussed elsewhere.In addition, because this work considers only single-beam studies (that is, only one beam is in operation and collisions are absent), events arising from beam-beam interactions, referred to as luminosity backgrounds, are also not discussed.Such backgrounds add an additional signal on top of those events present from the sources described below, and thus single-beam backgrounds are considered in isolation in this study.Future work is planned to include luminosity backgrounds.
We refer to backgrounds arising from interactions between particles in the same beam bunch as Touschek backgrounds.As a result of such interactions, beam particles gain (lose) energy, and thus propagate with an energy higher (lower) than the nominal bunch energy and with a correspondingly changed orbit.Particles whose energies are affected sufficiently will eventually collide with the beam pipe wall, causing a shower which may be incident on Belle II subdetectors.We expect the rate of this type of background to inversely proportional to the number of filled bunches in the Main Storage Ring (MR) and the bunch volume σ x × σ y × σ z , and the square of the beam current.
Another source of background arises when beam particles interact with residual free gas molecules inside the beam pipe, which we refer to as beam-gas backgrounds.Similar to the Touschek case, deviations from the nominal orbit and showers can result from such events.Because the likelihood of beam-gas interactions increases with the number of gas molecules and the cross section of beam particles and gas nuclei, we expect this background to be proportional to the pressure in the MR and also to the effective atomic number of the gas present in the ring.
Understanding the composition of beam-induced backgrounds thus provides an insight into accelerator conditions, and can be used to form extrapolations for future beam operation as the accelerator optics are brought closer to their design values.Furthermore, these extrapolations can be used to guide improvements to the accelerator, e.g., demonstrating the need for increased collimation to remove deviant particles at particular locations in the MR.
Here we report only the results of the effort carried out at Hiroshima University to develop a model of SuperKEKB beam-induced backgrounds using spatial information from the ECL, complementary to other similar studies reported elsewhere.In the future, we aim to include more mature results as a companion to other Belle II background analyses.Further, this work aims at providing a near-realtime monitoring system to be used by shifters to allow for fast feedback during beam operation.The Belle II detector, shown in cutaway in Fig. 1, is composed of a set of subdetectors situated about SuperKEKB's Interaction Point (IP).Individual subdetectors include a Vertex Detector situated close to the IP, a Central Drift Chamber for charged particle tracking, a Time of Propagation counter and focusing aerogel RICH for particle identification, an Electromagnetic Calorimeter, and an outer detector for detection of K L and muons.The Belle II subdetectors are described in detail in Ref. [5].

The Belle II ECL
The ECL is unique among these for its near-total coverage of the interaction region and for being "always on," allowing for measurements when other subdetectors must be powered down or gated.In particular, the ECL can record data during background studies and machine tests.The ECL is comprised of 8736 CsI(Tl) scintillator cells joined to PIN-PD photodetectors.These are divided into three broad regions: The Forward Endcap, Barrel, and Backward Endcap.A schematic of the ECL is shown in Fig. 2.

Procedure
In this study, we consider the spatial distribution of hits recorded in the ECL, in contrast to the studies based on detector hit rates in Refs.[2,3].We consider the distribution of background hits recorded during dedicated SuperKEKB single-beam studies, compared with Monte Carlo (MC) simulations of individual beam background types.Individual background MC distributions are fitted as independent parameters with the MINUIT fitting package in ROOT to form a beam background model complementary to those presented by other detectors.Individual fits are performed for the high and low energy rings (HER and LER, respectively).Individual endcaps and the barrel are also fitted separately.
Background sources are separated based on the best fits to spatial distributions of events within the ECL, with each source fitted as a floating parameter.
ECL data is recorded as clusters, in order to retain shower information, and only clusters with a total energy greater than 20 MeV are recorded.The MC is not clustered, but rather retains the position of each individual beam background hit.Therefore, in order to provide observables for comparison we consider in the data the position of the crystal with the largest individual energy deposition in each cluster as a proxy for the full event distribution.

Beam Background Data
Beam background signals were recorded during a series of dedicated single-beam background studies.As these studies have only one beam, backgrounds incident on the detectors come solely from a single beam.A summary of studies considered for this work is listed in Table 1.interaction region have their phase-space positions recorded and passed to GEANT4 [7] for more granular simulation of their passage through the detectors and digitization.

Results
Distributions for each of the three individual sections of the ECL were considered separately; in Table 2 and Figs. 3 and 4, we show the results for distributions of hits for the HER (LER) in the Forward (Backward) Endcap.The choice of different endcap is made based on the direction of beam propagation, as the endcap corresponding to the forward-going direction of each beam provides the best spatial information.Figure 3 shows the results of fits of individual background components to the recorded data distribution on the Forward Endcap for the dedicated HER studies, while Fig. 4 presents the results for the corresponding LER data.In both the HER and LER distributions, the spatial fit is shown to have a mismatch with data at the center and outer edge regions, while the central areas are better modeled.Future analyses will seek to determine the nature and cause of this discrepancy.showing the relative proportions of backgrounds resulting from Touschek, Beam-gas Coulomb, and Beam-gas Bremsstrahlung interactions.

Conclusion
Results obtained from fits shown in the previous section indicate discrepancies between recorded data and MC simulation at the edges of the endcap distribution, indicating a possible area of improvement for future MC.In particular, the inner and outer edges of each beam's respective endcap show discrepancies between the two distributions, while the center region of both are relatively well-modeled.Further work is needed to find the source of these discrepancies: similar fits of single-beam studies taken at other times during Belle II's running may be used to determine if this is a systematic effect, and a closer examination of beam conditions, particularly injection vs. storage period distributions, may also be useful in determining the cause of this discrepancy.
Comparison of the contributions of individual background types with previous results demonstrates only first-order agreement.Further work, as outlined above, may serve to bring these results closer in line with other detectors as the spatial discrepancies are resolved; once this work has thus matured, we also intend to combine it with other complementary analyses for the overall background measurement and reduction effort.With this and further work, we hope to make a near-realtime system for monitoring beam backgrounds at SuperKEKB which can be used during or shortly after beam operation to aid in beam operation.

Figure 1 .
Figure 1.Cutaway of Belle II, showing the individual subdetectors and the location of the electron and positron beam lines.Figure 2. Layout of the Belle II ECL

Figure 2 .
Figure 1.Cutaway of Belle II, showing the individual subdetectors and the location of the electron and positron beam lines.Figure 2. Layout of the Belle II ECL

Figure 3 .
Figure 3. Distribution of Monte Carlo beam backgrounds fitted to recorded HER data in the Forward Endcap of the ECL.The horizontal axis represents numbered crystal positions, with the inner (outer) portion of the endcap at left (right).

Figure 4 .
Figure 4. Distribution of Monte Carlo beam backgrounds fitted to recorded LER data in the Backward Endcap of the ECL.The horizontal axis represents numbered crystal positions, with the inner (outer) portion of the endcap at left (right).
Fits of background MC distributions produced for individual background types to recorded data in the ECL were performed for studies of individual beams in both the HER and LER at SuperKEKB.Using this data, we developed a spatial model of single-beam backgrounds 14th International Particle Accelerator Conference Journal of Physics: Conference Series 2687 (2024) 022008

Table 1 .
Summary of Single-Beam Studies under Consideration.HER (LER) refer to the high (low) energy beam rings.
[6]m background MC is generated separately for Touschek and beam-gas backgrounds, with beam-gas being further divided into Coulomb and Bremsstrahlung components, to better model individual physical processes involved in beam-gas scattering.Each distribution is first generated in SAD[6]with only the relevant physical process enabled; beam particles which enter the

Table 2 .
Summary of Fits of Background Distributions.Values for background components represent their percentage contribution to the entire distribution.